Your box is in the mail. You have three to five business days and nothing to do but think about what you're about to build. This is actually the most valuable time in the whole project — and there is something specific you should be doing with it.
Open a browser. Go to claude.ai or chatgpt.com. Get an account.
You are going to hear a lot of opinions about which AI service is best. Those opinions change weekly and are mostly beside the point. What matters is that you have access to one and that you learn to use it well. A paid subscription is worth it. This is one of the few places in this book where a subscription gets a genuine recommendation — because unlike a managed hosting service or a branded development tool, an AI assistant is not reselling you something you could own outright. It is a capability that does not exist any other way at this quality level. It is worth the money.
Now. Here is what most people get wrong about AI.
They treat it like a search engine. They type a few keywords, they get an answer, they copy and paste it, they move on. And then they wonder why the results are mediocre.
The AI was trained on language. Specifically, on good language — complete sentences, proper grammar, clear punctuation, ideas expressed precisely enough to be understood. The quality of what you get back is directly proportional to the quality of what you put in. Your junior high English teacher was preparing you for this moment and neither of you knew it.
Use full sentences. Use proper punctuation. Say what you actually mean. Use a browser with a spell checker that works in the text input area. Describe your situation completely before you ask your question. The AI cannot see your screen, does not know your setup, and has never met your server. Tell it everything that's relevant. Then ask.
This is not a search engine. This is a conversation. Have it like one.
Now here is the most useful thing you can do while you are waiting for that box to arrive.
Have the AI help you build a skills document.
A skills document is a description of your system — what you have, how it is configured, what you are trying to accomplish, what conventions you follow. Once you have one, you bring it into every AI conversation about your project. Suddenly the AI is not giving you generic Drupal advice. It is giving you advice about your system. It knows your stack, your server, your setup. It becomes a colleague who knows the codebase rather than a stranger you have to brief from scratch every time.
You do not have to write this document yourself. In fact you shouldn't — not alone. Start a conversation. Tell the AI what you are building. Describe what you are thinking. Let it ask you questions. Correct it when it gets something wrong. Adjust until you both agree on what is being built. Then say: "I want to build a skills document based on everything we've discussed. Can you write it for me?"
That document will make everything that follows easier.
Once your box arrives and your system is running, your AI assistant becomes genuinely useful in three specific ways. It can walk you through command line instructions and explain what each command does before you run it. It can help you write and debug code. And it is an excellent troubleshooter — paste an error message into the conversation, describe what you were doing when it appeared, and you will almost always get useful direction.
Almost always.
Here is the part that does not change regardless of how good the AI gets: you are responsible for every command you run and every line of code you deploy. The AI does not have root access to your server. You do. The AI is making an educated suggestion based on what you told it — and what you told it may not be the whole picture. Read every command before you run it. If you don't understand what it does, ask. "What will this command change on my system?" is not a stupid question. It is the right question. The AI got it wrong sometimes. When that happens and you ran the command anyway, that is still on you.
This is not a reason to distrust the tool. It is a reason to use it like a professional — which means maintaining your own judgment at every step. The AI is the knowledgeable colleague who suggests the approach. You are the one who decides whether to take the advice.
That is how professional judgment develops. Conversation by conversation. Command by command.